Network Working Group J. Welch
Request for Comments: 4445 IneoQuest Technologies
Category: Informational J. Clark
Cisco Systems
April 2006
A Proposed Media Delivery Index (MDI)
Status of This Memo
This memo provides information for the Internet community. It does
not specify an Internet standard of any kind. Distribution of this
memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society (2006).
IESG Note
This RFC is not a candidate for any level of Internet Standard.
There are IETF standards which are highly applicable to the space
defined by this document as its applicability, in particular, RFCs
3393 and 3611, and there is ongoing IETF work in these areas as well.
The IETF also notes that the decision to publish this RFC is not
based on IETF review for such things as security, congestion control,
MIB fitness, or inappropriate interaction with deployed protocols.
The RFC Editor has chosen to publish this document at its discretion.
Readers of this document should exercise caution in evaluating its
value for implementation and deployment. See RFC 3932 for more
information.
Abstract
This memo defines a Media Delivery Index (MDI) measurement that can
be used as a diagnostic tool or a quality indicator for monitoring a
network intended to deliver applications such as streaming media,
MPEG video, Voice over IP, or other information sensitive to arrival
time and packet loss. It provides an indication of traffic jitter, a
measure of deviation from nominal flow rates, and a data loss
at-a-glance measure for a particular flow. For instance, the MDI may
be used as a reference in characterizing and comparing networks
carrying UDP streaming media.
The MDI measurement defined in this memo is intended for Information
only.
Welch & Clark Informational [Page 1]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
1. Introduction
There has been considerable progress over the last several years in
the development of methods to provide for Quality of Service (QoS)
over packet-switched networks to improve the delivery of streaming
media and other time-sensitive and packet-loss-sensitive applications
such as [i1], [i5], [i6], [i7]. QoS is required for many practical
networks involving applications such as video transport to assure the
availability of network bandwidth by providing upper limits on the
number of flows admitted to a network, as well as to bound the packet
jitter introduced by the network. These bounds are required to
dimension a receiver`s buffer to display the video properly in real
time without buffer overflow or underflow.
Now that large-scale implementations of such networks based on RSVP
and Diffserv are undergoing trials [i3] and being specified by major
service providers for the transport of streaming media such as MPEG
video [i4], there is a need to diagnose issues easily and to monitor
the real-time effectiveness of networks employing these QoS methods
or to assess whether they are required. Furthermore, due to the
significant installed base of legacy networks without QoS methods, a
delivery system`s transitional solution may be composed of networks
with and without these methods, thus increasing the difficulty in
characterizing the dynamic behavior of these networks.
The purpose of this memo is to describe a set of measurements that
can be used to derive a Media Delivery Index (MDI) that indicates the
instantaneous and longer-term behavior of networks carrying streaming
media such as MPEG video.
While this memo addresses monitoring MPEG Transport Stream (TS)
packets [i8] over UDP, the general approach is expected to be
applicable to other streaming media and protocols. The approach is
applicable to both constant and variable bit rate streams though the
variable bit rate case may be somewhat more difficult to calculate.
This document focuses on the constant bit rate case as the example to
describe the measurement, but as long as the dynamic bit rate of the
encoded stream can be determined (the "drain rate" as described below
in Section 3), then the MDI provides the measurement of network-
induced cumulative jitter. Suggestions and direction for calculation
of MDI for a variable bit rate encoded stream may be the subject of a
future document.
Network packet delivery time variation and various statistics to
characterize the same are described in a generic approach in [i10].
The approach is capable of being parameterized for various purposes
with the intent of defining a flexible, customizable definition that
can be applied to a wide range of applications and further
Welch & Clark Informational [Page 2]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
experimentation. Other approaches to characterizing jitter behavior
are also captured such as in [i12]. A wide-ranging report format
[i11] has been described to convey information including jitter for
use with the RTP Control Protocol (RTCP) [i12]. The MDI is instead
intended to specifically address the need for a scalable,
economical-to-compute metric that characterizes network impairments
that may be imposed on streaming media, independent of control plane
or measurement transport protocol or stream encapsulation protocol.
It is a targeted metric for use in production networks carrying large
numbers of streams for the purpose of monitoring the network quality
of the flows or for other applications intended to analyze large
numbers of streams susceptible to IP network device impairments. An
example application is the burgeoning deployments of Internet
Protocol Television (IPTV) by cable and telecommunication service
providers. As described below, MDI provides for a readily scalable
per-stream measure focused on loss and the cumulative effects of
jitter.
2. Media Delivery Index Overview
The MDI provides a relative indicator of needed buffer depths at the
consumer node due to packet jitter as well as an indication of lost
packets. By probing a streaming media service network at various
nodes and under varying load conditions, it is possible to quickly
identify devices or locales that introduce significant jitter or
packet loss to the packet stream. By monitoring a network
continuously, deviations from nominal jitter or loss behavior can be
used to indicate an impending or ongoing fault condition such as
excessive load. It is believed that the MDI provides the necessary
information to detect all network-induced impairments for streaming
video or voice-over-IP applications. Other parameters may be
required to troubleshoot and correct the impairments.
The MDI is updated at the termination of selected time intervals
spanning multiple packets that contain the streaming media (such as
transport stream packets in the MPEG-2 case). The Maximums and
Minimums of the MDI component values are captured over a measurement
time. The measurement time may range from just long enough to
capture an anticipated network anomaly during a troubleshooting
exercise to indefinitely long for a long-term monitoring or logging
application. The Maximums and Minimums may be obtained by sampling
the measurement with adequate frequency.
Welch & Clark Informational [Page 3]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
3. Media Delivery Index Components
The MDI consists of two components: the Delay Factor (DF) and the
Media Loss Rate (MLR).
3.1. Delay Factor
The Delay Factor is the maximum difference, observed at the end of
each media stream packet, between the arrival of media data and the
drain of media data. This assumes the drain rate is the nominal
constant traffic rate for constant bit rate streams or the piece-wise
computed traffic rate of variable rate media stream packet data. The
"drain rate" here refers to the payload media rate; e.g., for a
typical 3.75 Mb/s MPEG video Transport Stream (TS), the drain rate is
3.75 Mb/s -- the rate at which the payload is consumed (displayed) at
a decoding node. If, at the sample time, the number of bytes
received equals the number transmitted, the instantaneous flow rate
balance will be zero; however, the minimum DF will be a line packet's
worth of media data, as that is the minimum amount of data that must
be buffered.
The DF is the maximum observed value of the flow rate imbalance over
a calculation interval. This buffered media data in bytes is
expressed in terms of how long, in milliseconds, it would take to
drain (or fill) this data at the nominal traffic rate to obtain the
DF. Display of DF with a resolution of tenths of milliseconds is
recommended to provide adequate indication of stream variations for
monitoring and diagnostic applications for typical stream rates from
1 to 40 Mb/s. The DF value must be updated and displayed at the end
of a selected time interval. The selected time interval is chosen to
be long enough to sample a number of TS packets and will, therefore,
vary based on the nominal traffic rate. For typical stream rates of
1 Mb/s and up, an interval of 1 second provides a long enough sample
time and should be included for all implementations. The Delay
Factor indicates how long a data stream must be buffered (i.e.,
delayed) at its nominal bit rate to prevent packet loss. This time
may also be seen as a measure of the network latency that must be
induced from buffering, which is required to accommodate stream
jitter and prevent loss. The DF`s max and min over the measurement
period (multiple intervals) may also be displayed to show the worst
case arrival time deviation, or jitter, relative to the nominal
traffic rate in a measurement period. It provides a dynamic flow
rate balance indication with its max and min showing the worst
excursions from balance.
The Delay Factor gives a hint of the minimum size of the buffer
required at the next downstream node. As a stream progresses, the
variation of the Delay Factor indicates packet bunching or packet
Welch & Clark Informational [Page 4]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
gaps (jitter). Greater DF values also indicate that more network
latency is necessary to deliver a stream due to the need to pre-fill
a receive buffer before beginning the drain to guarantee no
underflow. The DF comprises a fixed part based on packet size and a
variable part based on the buffer utilization of the various network
component switch elements that comprise the switched network
infrastructure [i2].
To further detail the calculation of DF, consider a virtual buffer VB
used to buffer received packets of a stream. When a packet P(i)
arrives during a calculation interval, compute two VB values,
VB(i,pre) and VB(i,post), defined as:
VB(i,pre) = sum (Sj) - MR * Ti; where j=1..i-1
VB(i,post) = VB(i,pre) + Si
where Sj is the media payload size of the jth packet, Ti is the
relative time at which packet i arrives in the interval, and MR is
the nominal media rate.
VB(i,pre) is the Virtual Buffer size just before the arrival of P(i).
VB(i,post) is the Virtual Buffer size just after the arrival of P(i).
The initial condition of VB(0) = 0 is used at the beginning of each
measurement interval. A measurement interval is defined from just
after the time of arrival of the last packet during a nominal period
(typically 1 second) as mentioned above to the time just after the
arrival of the last packet of the next nominal period.
During a measurement interval, if k packets are received, then there
are 2*k+1 VB values used in deriving VB(max) and VB(min). After
determining VB(max) and VB(min) from the 2k+1 VB samples, DF for the
measurement interval is computed and displayed as:
DF = [VB(max) - VB(min)]/ MR
As noted above, a measurement interval of 1 second is typically used.
If no packets are received during an interval, the last DF calculated
during an interval in which packets did arrive is displayed. The
time of arrival of the last previous packet is always retained for
use in calculating VB when the next packet arrives (even if the time
of the last received packet spans measurement intervals). For the
first received measurement interval of a measurement period, no DF is
calculated; however, packet arrival times are recorded for use in
calculating VB during the following interval.
Welch & Clark Informational [Page 5]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
3.2. Media Loss Rate
The Media Loss Rate is the count of lost or out-of-order flow packets
over a selected time interval, where the flow packets are packets
carrying streaming application information. There may be zero or
more streaming packets in a single IP packet. For example, it is
common to carry seven 188 Byte MPEG Transport Stream packets in an IP
packet. In such a case, a single IP packet loss would result in 7
lost packets counted (if those 7 lost packets did not include null
packets). Including out-of-order packets is important, as many
stream consumer-type devices do not attempt to reorder packets that
are received out of order.
3.3. Media Delivery Index
Combining the Delay Factor and Media Loss Rate quantities for
presentation results in the following MDI:
DF:MLR
Where:
DF is the Delay Factor
MLR is the Media Loss Rate
At a receiving node, knowing its nominal drain bit rate, the DF`s max
indicates the size of buffer required to accommodate packet jitter.
Or, in terms of Leaky Bucket [i9] parameters, DF indicates bucket
size b, expressed in time to transmit bucket traffic b, at the given
nominal traffic rate, r.
3.4. MDI Application Examples
If a known, well-characterized receive node is separated from the
data source by unknown or less well-characterized nodes such as
intermediate switch nodes, the MDI measured at intermediate data
links provides a relative indication of the behavior of upstream
traffic flows. DF difference indications between one node and
another in a data stream for a given constant interval of calculation
can indicate local areas of traffic congestion or possibly
misconfigured QoS flow specification(s) leading to greater filling of
measurement point local device buffers, resultant flow rate
deviations, and possible data loss.
Welch & Clark Informational [Page 6]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
For a given MDI, if DF is high and/or the DF Max-Min captured over a
significant measurement period of multiple intervals is high, jitter
has been detected but the longer-term, average flow rate may be
nominal. This could be the result of a transient flow upset due to a
coincident traffic stream unrelated to the flow of interest causing
packet bunching. A high DF may cause downstream buffer overflow or
underflow or unacceptable latency even in the absence of lost data.
Due to transient network failures or DF excursions, packets may be
lost within the network. The MLR component of the MDI shows this
condition.
Through automated or manual flow detection and identification and
subsequent MDI calculations for real-time statistics on a flow, the
DF can indicate the dynamic deterioration or increasing burstiness of
a flow, which can be used to anticipate a developing network
operation problem such as transient oversubscription. Such
statistics can be obtained for flows within network switches using
available switch cpu resources due to the minimal computational
requirements needed for small numbers of flows. Statistics for all
flows present on, say, a gigabit Ethernet network, will likely
require dedicated hardware facilities, though these can be modest, as
buffer requirements and the required calculations per flow are
minimal. By equipping network switches with MDI measurements, flow
impairment issues can quickly be identified, localized, and
corrected. Until switches are so equipped with appropriate hardware
resources, dedicated hardware tools can provide supplemental switch
statistics by gaining access to switch flows via mirror ports, link
taps, or the like as a transition strategy.
The MDI figure can also be used to characterize a flow decoder's
acceptable performance. For example, an MPEG decoder could be
characterized as tolerating a flow with a given maximum DF and MLR
for acceptable display performance (acceptable on-screen artifacts).
Network conditions such as Interior Gateway Protocol (IGP)
reconvergence time then might also be included in the flow tolerance
DF resulting in a higher-quality user experience.
4. Summary
The MDI combines the Delay Factor, which indicates potential for
impending data loss, and Media Loss Rate as the indicator of lost
data. By monitoring the DF and MLR and their min and max excursions
over a measurement period and at multiple strategic locations in a
network, traffic congestion or device impairments may be detected and
isolated for a network carrying streaming media content.
Welch & Clark Informational [Page 7]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
5. Security Considerations
The measurements identified in this document do not directly affect
the security of a network or user. Actions taken in response to
these measurements that may affect the available bandwidth of the
network or the availability of a service is out of scope for this
document.
Performing the measurements described in this document only requires
examination of payload header information (such as MPEG transport
stream headers or RTP headers) to determine nominal stream bit rate
and sequence number information. Content may be encrypted without
affecting these measurements. Therefore, content privacy is not
expected to be a concern.
6. Informative References
[i1] Braden, R., Zhang, L., Berson, S., Herzog, S., and S. Jamin,
"Resource ReSerVation Protocol (RSVP) -- Version 1 Functional
Specification", RFC 2205, September 1997.
[i2] Partridge, C., "A Proposed Flow Specification", RFC 1363,
September 1992.
[i3] R. Fellman, `Hurdles to Overcome for Broadcast Quality Video
Delivery over IP` VidTranS 2002.
[i4] CableLabs `PacketCable Dynamic Quality-of-Service
Specification`, PKT-SP-DQOS-I06-030415, 2003.
[i5] Shenker, S., Partridge, C., and R. Guerin, "Specification of
Guaranteed Quality of Service", RFC 2212, September 1997.
[i6] Wroclawski, J., "Specification of the Controlled-Load Network
Element Service", RFC 2211, September 1997.
[i7] Braden, R., Clark, D., and S. Shenker, "Integrated Services in
the Internet Architecture: an Overview", RFC 1633, June 1994.
[i8] ISO/IEC 13818-1 (MPEG-2 Systems)
[i9] V. Raisanen, "Implementing Service Quality in IP Networks",
John Wiley & Sons Ltd., 2003.
[i10] Demichelis, C. and P. Chimento, "IP Packet Delay Variation
Metric for IP Performance Metrics (IPPM)", RFC 3393, November
2002.
Welch & Clark Informational [Page 8]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
[i11] Friedman, T., Caceres, R., and A. Clark, "RTP Control Protocol
Extended Reports (RTCP XR)", RFC 3611, November 2003.
[i12] Schulzrinne, H., Casner, S., Frederick, R., and V. Jacobson,
"RTP: A Transport Protocol for Real-Time Applications", STD 64,
RFC 3550, July 2003.
7. Acknowledgements
The authors gratefully acknowledge the contributions of Marc Todd and
Jesse Beeson of IneoQuest Technologies, Inc., Bill Trubey and John
Carlucci of Time Warner Cable, Nishith Sinha of Cox Communications,
Ken Chiquoine of SeaChange International, Phil Proulx of Bell Canada,
Dr Paul Stallard of TANDBERG Television, Gary Hughes of Broadbus
Technologies, Brad Medford of SBC Laboratories, John Roy of Adelphia
Communications, Cliff Mercer, PhD of Kasenna, Mathew Ho of Rogers
Cable, and Irl Duling of Optinel Systems for reviewing and evaluating
early versions of this document and implementations for MDI.
Authors' Addresses
James Welch
IneoQuest Technologies, Inc
170 Forbes Blvd
Mansfield, Massachusetts 02048
Phone: 508 618 0312
EMail: Jim.Welch@ineoquest.com
James Clark
Cisco Systems, Inc
500 Northridge Road
Suite 800
Atlanta, Georgia 30350
Phone: 678 352 2726
EMail: jiclark@cisco.com
Welch & Clark Informational [Page 9]
RFC 4445 A Proposed Media Delivery Index (MDI) April 2006
Full Copyright Statement
Copyright (C) The Internet Society (2006).
This document is subject to the rights, licenses and restrictions
contained in BCP 78 and at www.rfc-editor.org/copyright.html, and
except as set forth therein, the authors retain all their rights.
This document and the information contained herein are provided on an
"AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
Intellectual Property
The IETF takes no position regarding the validity or scope of any
Intellectual Property Rights or other rights that might be claimed to
pertain to the implementation or use of the technology described in
this document or the extent to which any license under such rights
might or might not be available; nor does it represent that it has
made any independent effort to identify any such rights. Information
on the procedures with respect to rights in RFC documents can be
found in BCP 78 and BCP 79.
Copies of IPR disclosures made to the IETF Secretariat and any
assurances of licenses to be made available, or the result of an
attempt made to obtain a general license or permission for the use of
such proprietary rights by implementers or users of this
specification can be obtained from the IETF on-line IPR repository at
http://www.ietf.org/ipr.
The IETF invites any interested party to bring to its attention any
copyrights, patents or patent applications, or other proprietary
rights that may cover technology that may be required to implement
this standard. Please address the information to the IETF at
ietf-ipr@ietf.org.
Acknowledgement
Funding for the RFC Editor function is provided by the IETF
Administrative Support Activity (IASA).
Welch & Clark Informational [Page 10]